GaWC Research Bulletin 177

GaWC logo
  Gateways into GaWC

This Research Bulletin has been published in PJ Taylor, B Derudder, P Saey and F Witlox (eds) (2007) Cities in Globalization: Practices, Policies and Theories London: Routledge, pp. 287-297.

Please refer to the published version when quoting the paper.


Cities within Spaces of Flows: Theses for a Materialist Understanding of the External Relations of Cities

P.J. Taylor


In the conference from which this volume is derived, I used a format of 48 theses with the intension of producing a 'punchy' presentation. But there were two more basic reasons for this choice of mode of argument. First, I was at an early stage in my thinking on the subject matter and laying out a skeletal argument was both necessary and at the limit of my thoughts. Second and related to the latter, the subject matter I had chosen encompassed a scope of ideas that was unusually broad ranging: the method of short, sharp statements as theses was a way of compressing the tentative argument into a manageable whole for the audience. All these conditions still exist and thus I have continued with this format in this chapter. However, as well as revising, extending and reorganising the theses as a result of further thinking, I have added this prologue to indicate how I have come to the subject matter at hand.

The spark for the train of thought that are the theses below came from observing the same basic statement from two outstanding urban theorists writing half a century apart: in 1921 Max Weber (1958, 65) argued that the city was is not simply 'a large locality' and in 1969 Jane Jacobs (1970, 129) argued that 'a city is not a large town' - I used the latter quote as my title for the conference presentation. Of course, these theorists subsequently diverge in their respective paths to city definition: Weber focuses on functions, Jacobs understands cities as a process (Jacobs, 1970, 50). Move on another quarter century and we find Jacobs echoed in Castells' (1996, 386) influential work on network society: 'the global city is ... a process'. Although the latter is based upon another different theory drawing on ideas of global cities atop a world urban hierarchy, Jacobs and Castells share a similar relational view of cities. The materialist understanding I try and develop below attempts to bring together key aspects of Jacobs' and Castells' ideas on cities and space.

But I do not divorce their ideas from traditional geographical concerns for cities and space. Going back as far as the geographer-anarchist Elisée Reclus (1895), I have long been intrigued by his postulating a flat (i.e. isotropic) plain to derive a regular geometric pattern of cities long before Christaller (1966) published his treatise on central place theory in 1933. On a superficial reading of Reclus I had thought he presaged the central place model but he actually proposes a quite different basis for the distribution of cities 'regulated ... by the step of the traveller' (p. 251). Here is a model of the spacing of cities not as central places but as destinations. I draw directly on this possibility of different spatial orderings of urban places to develop theses on how cities differ from towns. Thus central place theory appears in my theses but I am able to curtail its extrapolation into 'city systems' that are 'national urban hierarchies'. This requires a means of separating economic space production from economic space production. For this I go to another part of Jacobs' oeuvre that I relate to Arrighi's (1994) emphasis on separating profit-orientated business enterprises from power-orientated government organization in the rise of the capitalist world-economy. Commonly 'political economy' approaches emphasise the need to integrate politics and economics and this is achieved through the myth of the territorial coincidence of 'national economy' with nation-state. But intriguingly, Arrighi (1994, 109-27) shows that in the initial manifestation of a modern world–system in the sixteenth century, the formative powers were not geographically coincident: the business enterprises of Genoa combined symbiotically with the protective power of Castile in a relation that is difficult to comprehend through modern lenses. It was by thinking through the implications of this early modern city/state organisation that I was able to link it to Jacobs' (1992) ideas on alternative ways of making a living, one essentially economic the other political. Thus the theses that follow respect a separation between the economic and the political: the argument is against their study through integration a la political economy, but rather through their inter-relations as distinctive social activities.

As will be gathered from above, the theses draw on ideas covering many decades. This may seem an unfashionable approach to take; there is a tendency today to use a set of ideas for a short while before moving on to the next set that seem to be more attractive in a sort of 'up-to-date' way. But Berman (1984) has warned us that in any realm of social activity, the fashion-conscious are always in danger of becoming objects of modernity rather than subjects in modernity. For social scientists, in particular, being objects rather than subjects must be totally debilitating. For instance, 'modern fashion police' should have objectified my work by forcing abandonment of world-systems analysis sometime in the 1980s: in current undergraduate textbooks that my students use for understanding 'development', world-systems analysis is conflated with the 'dependency approach' which superseded the 'modernization approach' in the 1970s but has itself been superseded by several new inventions of development since. In the theses below I keep to 'old-fashioned' world-systems ideas since my reasons for adopting this approach (relating to continuing world poverty) remain as strong as ever. The hurry to flip between discourses may be intellectually stimulating but I fear it loses key insights in dismissal of 'old ideas'. My argument is not so much about 'building paradigms' or avoiding perennial 'reinventing the wheel', it is simply a plea to understand that profound thinkers are not just of historical interest – 'when were their ideas superseded?' is not necessarily the best question to ask. Rather, where the same ideas appear in several contexts it might just mean that we are in the presence of some fundamental notions, perhaps even truly profound ideas, that need to be cherished and used rather than discarded at the behest of the latest fashion.

I should make it clear that my theses are not a summary of past ideas, rather I try to use those ideas and juxtaposition them in new ways and with new ideas to create fresh views of inter-city relations. I have organised the theses into four sets of ideas that constitute my argument: first, the foundation premises underpinning my position; second, I lay out my theses on city as process and differentiate 'city-ness from 'town-ness'; third and fourth, I develop what are largely corollary theses that are transhistorical and specifically modern respectively.


I. Cities

  1. Cities have been in existence for some thousands of years across many regions of the world; an understanding of these social entities, therefore, must be trans-historical and geo-graphical in nature.

  2. Cities always come in packs, never alone; it follows that urban theories and theories of 'the city' that deal with these social entities as singular are inherently inadequate.

  3. Cities must be studied relationally as inter-city relations; case studies, comparative studies and even 'throughput' studies (for instance, Harvey's (1982) circulation/fixed capital; Lebfevre's (1996) particle/wave analog; Cox and Mair's (1988) local coalitions capturing flows, and Massey's (1993) progressive place) are inadequate conceptual bases for understanding cities.

II. Cities and States

  1. Cities are social entities primarily created by economic processes; it is the study of work carried out in and through cities that will provide the necessary starting point for a materialist theory of cities.

  2. States are social entities primarily created by political processes; behind the work of states there always lies the threat of force.

  3. Relations between cities and states constitute an intertwined geohistory of cooperation and competition concerning wealth and protection; ultimately cities cannot be understood without exploring their relations with states for which Jacobs (1992) provides a basic starting point.

III. Strategies of Social Reproduction and Their 'Moral Syndromes'

  1. Following Jacobs (1992), there are only two strategies of social reproduction: raiding/taking/protecting that leads to states, and trading/making/producing that leads to cities. Over millennia the activities towards these strategies have each created a distinctive ethics which Jacobs calls their moral syndromes.

  2. In the guardian moral syndrome the primary ethical precept of the agents is loyalty, the behavioural premise is a zero-sum game, and the key institutions are force and balance of power. For instance, a breakdown in loyalty causes catastrophe for an army, and its territorial losses will be mirrored by its enemy's territorial gains.

  3. In the commercial moral syndrome the primary ethical precept is honesty, the behavioural premise is a win-win scenario (what Hicks (1969, 44) calls the 'around advantage' of voluntary trade), and the key institutions are contracts and markets. For instance, a bank can only operate on the basis of honesty for the mutual benefit of lenders and borrowers, and a breakdown in this trust causes catastrophe for bank and customers.

IV. Spaces of Social Reproduction

  1. Each moral syndrome underpins behaviour that makes and uses particular forms of space that are Castells' (1996) spaces of places and spaces of flows.

  2. The dominant space of guardian activities is a space of places, territories to be won and protected (examples range from hunter/gatherer territories to the many varieties of states). The familiar mosaic map of modern nation-states is a particularly powerful contemporary space of places.

  3. The dominant space of commercial activities is a space of flows, movement of commodities for production and trade (examples range from ancient trading networks to many varieties of financial markets). Relations between 'global cities' as 'international financial centres' are a particularly powerful contemporary space of flows.


V. Cities, Towns and 'Urban'

  1. A city is not a large town; neither settlement type can be defined by the simple attribute of demographic size because both are best conceptualised as processes. This is the key thesis of the argument.

  2. The concept of 'urban' as contrasted to rural is a chaotic conception because it conflates town and city. Note that urbanization, a demographic territorial attribute, is typically used as a derogatory anti-city concept indicating a threat to a space of place that is rural. If a territorial term is required, it might be best to replace urbanization by de-ruralization.

  3. 'Urban' implies a concentration of population – a node in a space of flows - but the spaces created by towns and cities are inherently different; the contrast is between local hierarchies and non-local networks respectively.

VI. Towns and Christaller's Central Place Theory

  1. Towns are constituted by the processes that generate 'urban hinterlands'; the town-hinterland relationship may vary geohistorically but in essence towns are based upon local relations as described by phrases such as 'country towns' and 'market towns'.

  2. Christaller's (1966) central place theory defines such relations and shows formally how towns are arranged evenly and hierarchically.

  3. Guardian thinking on cities has used central place theory to define 'urban systems' as constituting national urban hierarchies that can be 'regulated' (Bourne 1976). But this is to incorporate leading national cities in this town-making process, with hinterlands becoming 'regional' (and ultimately 'national' for the top city) rather than strictly local.

VII. Beyond Town-ness

  1. A key limitation of central place theory is that it treats cities as large towns in violation of our thesis (V(i)). I am not arguing that cities do not have local hinterlands but that they are not to be defined as simply the towns with the largest hinterlands. It is necessary to consider cities beyond this one particular relational theory.

  2. A major advantage of defining both towns and cities as processes (thesis V(i)) is that they 'town-ness' and 'city-ness' can operate simultaneously through the same settlement. Thus it is that cities service a hinterland and in doing so they exhibit their 'town-ness'.

  3. In addition, of course, cities also exhibit a primary 'city-ness', a completely different process. This is the reason why central place theory's prediction of spatial regularity tends to work much better at lower levels of the hierarchy (towns) than upper echelons where dominating city-ness processes tend to override central places imperatives of town-ness.

VIII. Cities and Jacobs' Import Replacement Theory

  1. Jacobs (1970; 1984) defines a city as a settlement that has experienced one or more rapid spurts of new work, adding to old work, and thus producing a more complex and diverse economy than a town.

  2. But cities do not arise alone; the most important new work is import replacement whereby a city begins to produce what it previously imported from other cities in a network of cities.

  3. This is not a zero-sum game because the generation of new exports to replace lost exports, and conversion of import replacements to become new exports, both lead to a dynamic city network based upon evolving mutualities in an overall expansion of economic life.

IX. City-ness and Town-ness Contrasted

  1. City-ness is net-work, the development of a network of cities; town-ness is hinter-work, the development of a hierarchy of towns and their 'service areas'.

  2. Being part of a hierarchy is very different from being a member of a network. Town-ness is to know your place of dependence on settlements higher in the hierarchy. The import replacement of city-ness is a declaration of economic autonomy from such relations through the mutualities of a city network.

  3. Since these are two completely different processes it follows that there can be no simple settlement evolution theory – farm, village, town, city – with the latter as the most 'developed' settlement. This has important geohistorical implications because it means that towns and cities have different geohistories creating different spaces.


X. Two Processes, two 'Urban Revolutions'

  1. In the orthodox materialist interpretation of the origins of cities – there is only one urban revolution (Childe 1950) - 'the' urban revolution based upon the agricultural surpluses consequent upon the expansion of farming into productive riverine environments (notably Mesopotamia); but if there are two distinctive 'urban' processes, if follows that there should two such 'revolutions, one for towns and one for cities.

  2. The traditional agricultural explanation of the 'urban revolution' can now be interpreted as a creation of town-ness; increased yields per area unit provided local agricultural surpluses enabling some settlements to become towns by carving out hinterlands, and with a settlement hierarchy process no doubt encouraged by guardian agents through political control of food storage. Of course, the riverine environment was also highly conducive to trade expansion and it is through this process that cities will have developed.

  3. But trading long predates, not only the traditional 'urban revolution', but also agriculture itself. Hence, Jacobs' (1970) controversial thesis of 'cities first, then agriculture' is relevant here: the 'city revolution' may well be before the Neolithic, hunters and gatherers becoming long distance traders prior to the emergence of local farmers. Thus cities predate towns.

XI. Cities and Trade

  1. The existence of trade, however, does not inevitable mean a network with nodes as potential cities - frontier trade, relay trade, and peddling trade do not require cities. Diasporic trade, on the other hand, is a sure sign of city-ness; this results from communities of permanent 'foreign' traders acting as factors, agents, facilitators and knowledge sources for their kith and kin from other cities within the network.

  2. Trade is stimulated by areal differentiation; initially largely natural (resource) differences would be the stimulus, but social (production) differences have come to be more important; unlike central place theory such processes do not lead to spatial regularity – natural differences will presumably create erratic city locations, social differences for complementary trading may well create regionally clustered patterns of cities.

  3. Creators of new net–work, foreign agents, convert initial entrepôt (warehouse functions) into vibrant cosmopolitan cities by interlocking cities into a network.

XII. Net-work

  1. Net-work occurs in two basic forms: in Blaut's (1976) terms producing a commodity for market by change-of-form (manufacture) or by change-of-place (exchange).

  2. These create two spaces of flows: in Jacobs (1970) terms the 'little movements in the hubs' (intra-city relations – small-scale economic clusters) and the 'great wheels of economic life' (inter-city relations – large-scale economic clusters).

  3. These both contribute to the expansion of economic life as externalities, increased returns through non-market mechanisms (Hicks 1969). Internal 'cluster externalities' and external 'network externalities' are critical to vibrant net-work and consequent economic expansion.

XIII. Interlocking Networks in Time

  1. City networks are dynamic but not so all cities, all of the time; cities in networks rise and fall relative to each other.

  2. One of the most important properties of city networks is that at any one time while some cities are relatively economically stagnant (producing no new work), others continue to generate wealth – producing new net-work that provides the conditions for the stagnant to become dynamic again through a new round of import substitution.

  3. City networks exhibit historical layers of new work investment patterns across cities (Massey 1984); it is this dynamic geography - never reducing to no new work - that maintains the expansion of economic life.

XIV. City States and City Leagues

  1. Part of city wealth creation is used to buy protection from states but this guardian activity has also been provided by the cities themselves through both city-states and city leagues; territorial states are not necessary for wealth creation but in practice protection has usually been outsourced to such states.

  2. The Mesopotamian 'urban' revolution provides the first example of 'city-states', independent cities with their own governments; note, however, that this political independence – developing its own guardians - is not a requirement for city success, what is needed is a degree of economic autonomy to participate fully in the interlocking network.

  3. City leagues were a common medieval European political structure that aided in creation of city group monopolies but again, these city guardian activities are not necessary for city success; there were many ad hoc arrangements between cities and their merchants across the world that provided the required city autonomy to prosper in interlocking networks. The lex mercatoria as a legal system for merchants by merchants remains the classic example of not outsourcing to guardians.


XV. Cities in Nation-states

  1. Modern nation-states are unique among states for their attempts to build bounded 'national economies' (initially by emulating city 'mercantilist' policies (Heckscher 1955)); the result has been economically arbitrary spatial amalgams of city economies (Jacobs 1984).

  2. These attempts to tame economic spaces of flows through curtailing them in political spaces of places has succeeded in distorting the former to the degree that 'national urban systems' are commonly identified (Bourne and Simmons 1978); in reality, bounded economic protectionism has only operated successfully in large states where there are enough cities to constitute a viable national interlocking sub-network (e.g. USA, Germany, Japan).

  3. The general affect of this 'nationalization' of (cosmopolitan) cities has been to favour a few leading cities, often just the capital city, so that the patch of the world-economy controlled by the state exhibits severe hierarchical city relations (superficially conforming to central place predictions).

XVI. World City Network

  1. Today there is an interlocking network of cities wherein the diasporic feature (ex-pats) has been reorganised because of communication-computing technologies that provide instantaneous world-wide communication; this facilitates new capacities for management and servicing of multilocational production world-wide.

  2. This interlocking network is most clearly expressed through the new net-work associated with advanced producer service firms in professional, creative and financial work; major firms have world-wide office networks to provide a 'global seamless' service to clients (Taylor 2004).

  3. This new layer of investment (offices in the ubiquitous world city tower blocks) that creates new work is the world city network, a critical component of economic expansion in contemporary globalization.

XVII. Cities in Globalization

  1. Nation-states have not disappeared with globalization, rather they are key components of a much more complex social organization; in terms of inter-city relations this may be reflected in accentuating 'national hierarchies' as one city captures the 'gateway' function between the 'national' economy and world-economy: the classic case is Osaka's relative decline because of its 'Tokyo problem' (Hill and Fujita 1995) - other 'second city losers' may be Melbourne, Rio de Janeiro, Lyon, St Petersburg, and all British cities that are not London).

  2. However, contemporary intertwining of cities and states is much more complicated than this 'gateway' outcome suggests because the enabling distance reducing/eliminating technologies allow complex mixtures of centralization and decentralization.

  3. 'Gateways' can be easily by-passed using new communication technologies and therefore it is too soon to predict simple hierarchical enhancement since all cities are partially 'released' from national bounds through global networking.

XVIII. Cities and the World-Economy

  1. Although the world-economy's core-periphery pattern is usually described and depicted as combinations of nation-states. However, since this is an economic process it is, perhaps better not to describe it through guardian institutions, but through city economies in their networks: Wallerstein's description of core processes can be interpreted as city-making processes (both produce spatially clustered 'high tech' outcomes), and peripheral processes – the development of underdevelopment – can be similarly interpreted as Jacob's (1984) malign projection of core-city economic processes.

  2. In the contemporary world-economy, therefore, the core is defined by the processes of new work that are constituting the world city network, and the periphery is the rest of the world beyond the world city network.

  3. The semi-periphery is defined in Wallersteinian terms as locales where core and periphery processes are approximately balanced; these are cities in the erstwhile 'third world' that are now part of the world city network but are also 'mega-cities' (pernicious population 'town' growth that is a periphery process). This is what makes cities such as Sao Paulo, Mexico City, Mumbai, Johannesburg, and Bangkok among the most interesting settlements in the first decades of the twenty first century.


Arrighi, G. (1994) The Long Twentieth Century. London: Verso

Blaut, J. M. (1976) 'Where was capitalism born?', Antipode 8(2), 1-11

Bourne, L. S. (1975) Urban Systems: Strategies for Regulation. Oxford: Clarendon

Bourne, L. S. and Simmons, J. W. (eds) (1978) Systems of Cities. New York: Oxford University Press

Castells, M. (1996) The Rise of Network Society. Oxford: Blackwell

Childe, V. G. (1950) 'The urban revolution', Town Planning Review 21 (1), 3-17

Christaller, W. (1966) Central Places in Southern Germany (trans. C. W. Baskin). Englewood Cliffs, NJ: Prentice Hall

Cox, K. R. and Mair, A. (1988) 'Locality and community in the politics of local economic development', Annals, Association of American Geographers 78, 307-25

Harvey, D. (1982) The Limits to Capital. Oxford: Blackwell

Heckscher, E. F. (1955) Mercantilism (second ed.). London: George Allen and Unwin

Hicks, J. (1969) A Theory of Economic History. Oxford: Clarendon

Hill, R. C. and Fujita, K. (1995) 'Osaka's Tokyo problem', International Journal of Urban and Regional Research 19, 181-91

Jacobs, J. (1970) The Economy of Cities. New York: Vintage

Jacobs, J. (1984) Cities and the Wealth of Nations. New York: Vintage

Jacobs, J. (1992) Systems of Survival. New York: Vintage

Lefebvre, H. (1996) Writings on Cities (trans. E. Kofman and E. Lebas). Oxford: Blackwell

Massey, D. (1984) Spatial Divisions of Labour. London: Macmillan

Massey, D. (1993) 'Power geometry and a progressive sense of place' in J. Bird et al (eds) Mapping the Futures. London: Routledge

Reclus, E. (1895) 'The evolution of cities', The Contemporary Review 67(2), 246-64

Taylor, P. J. (2004) World City Network: a Global Urban Analysis. London: Routledge

Weber, M. (1978) The City (trans. D. Martindale and G Neuwirth). New York: Free Press

Edited and posted on the web on 20th September 2005

Note: This Research Bulletin has been published in PJ Taylor, B Derudder, P Saey and F Witlox (eds) (2007) Cities in Globalization: Practices, Policies and Theories London: Routledge, pp. 287-297